Roche expands Dyno gene therapy collab, bets up to $1B-plus that AI can help design better delivery tools

After forming a gene therapy partnership with Dyno Therapeutics in 2020, Roche is back for more.

In a new deal potentially worth more than $1 billion, Roche is paying Dyno $50 million upfront to design novel adeno-associated virus (AAV) vectors with “improved functional properties” as delivery tools for gene therapies, Dyno said Thursday.

Roche is looking to use Dyno’s technologies to target neurological diseases, a big focus at the Swiss pharma, with multiple sclerosis blockbuster Ocrevus serving as its best-selling asset.

Dyno’s platform incorporates artificial intelligence and high-throughput in vivo data to help engineer and optimize AAV capsids. The Massachusetts biotech boasts the ability to measure the in vivo function of new sequences to the tune of billions in a month.

AAVs are widely accepted vehicles to deliver gene therapies, including in Roche’s Luxturna for a rare eye disease and Novartis’ Zolgensma for spinal muscular atrophy, a neurological disorder.

Existing AAV vectors based on naturally occurring viruses have various shortfalls. Some people may have preexisting immunity against an AAV, rendering the gene therapy it carries ineffective. Liver toxicity, poor tissue targeting and difficulty in manufacturing are also major problems with existing options.

Dyno believes man-made AAVs developed with its platform can improve tissue targeting, immune-evasion and scalability.

The latest deal builds on an initial collaboration Roche signed with Dyno in 2020 to develop central nervous system and liver-directed gene therapies. That first deal could exceed $1.8 billion in clinical and sales milestones. The new tie-up “provides Roche further access” to Dyno’s platform, according to the biotech.

“Our previous collaboration with Dyno Therapeutics gives us great confidence to increase our investment in therapeutic gene delivery, to support our neurological disease portfolio,” Roche’s newly minted head of corporate business development, Boris Zaïtra, said in a statement Thursday.

Dyno also counts Sarepta Therapeutics and Astellas among its partners.

Roche made a big commitment to gene therapies with its $4.3 billion acquisition of Luxturna maker Spark Therapeutics in 2019. Yet, five years later, Luxturna is still Spark’s lone commercial product. Earlier this year, Roche also ditched a gene therapy candidate for the neuromuscular disorder Pompe disease after analyzing the treatment landscape.

The lack of progress at Spark didn’t stop Roche from investing further in gene therapies. Besides Dyno, Roche has over the years teamed with Avista Therapeutics also on novel AAV capsids, with SpliceBio to work on a new treatment for an inherited retinal disease and with Sarepta on the Duchenne muscular dystrophy med Elevidys.

Meanwhile, some other large pharma companies have been shifting away from AAVs. For example, in a major pivot unveiled last year, Takeda ended its early-stage discovery and preclinical work on AAV-based gene therapies. Similarly, Pfizer effectively cut internal research efforts in viral-based gene therapies and last year offloaded a portfolio of preclinical gene therapy programs and related technologies to AstraZeneca’s rare disease unit Alexion.

The latest Dyno deal also follows several setbacks Roche has suffered in the neurology field. Besides the termination of the Pompe gene therapy program, Roche has recently returned the rights to UCB’s anti-tau antibody bepranemab in Alzheimer’s disease. And let’s not forget the surprise high-profile failure of the anti-amyloid antibody gantenerumab. In addition, anti-IL-6 drug Enspryng also came up short earlier this year in generalized myasthenia gravis, a neuromuscular autoimmune disorder.